Summary of Final Report, BB-305

June 2000

SEAL SLAB/PILE INTERFACE BOND

PROBLEM STATEMENT
Bridge superstructures are commonly supported on pile foundations, in which case, the
superstructure loads are transferred to a reinforced concrete pile cap that ties the supporting
individual piles into a complete structural unit. The piles themselves may be made of reinforced
concrete, prestressed concrete or steel.
When the required elevation of the pile cap is below the existing water table (i.e., in the case of
excavations or over-water bridges), de-watering of the foundation area must occur to allow for the
accurate placement of the reinforcing steel and pouring of the concrete. If global de-watering is not
possible because of induced adverse ground settlement or is impractical because of the proximity of
a body of water, a cofferdam must be employed.
A cofferdam is a temporary structure usually constructed of thin sheet piles that interconnect to form
a water-tight perimeter. Typically, the cofferdam sheet piles are installed first. Then, using a template
which locates the pile positions, the piles are driven. Both of the processes are conducted in the
saturated or submerged conditions. At this stage of the construction, a cast-in-place concrete (CIP)
seal slab is poured at the bottom of the cofferdam with the use of a tremie. This seals the bottom of
the cofferdam, preventing the seepage of water, and completes the coffer cell. By design, the top
elevation of the seal slab is the base elevation for the structural pile cap. The sheet piles and the seal
slab provide the formwork for the reinforced concrete pile cap.
As the function of the seal slab is primarily to provide a dry working surface, its design is relatively
unsophisticated. Under current design guidelines it is an unreinforced concrete slab with its depth
selected so that its weight largely offsets maximum uplift forces (for the maximum safe elevation of
water outside the cofferdam when completely dewatered). Allowance for interface bond between the
seal slab and the piles is minimal--allowable interface bond is 40 psi for concrete piles and 5 psi for
steel piles.

OBJECTIVES
The objective of this study was to evaluate the interface bond between a cast-in-place seal slab and
prestressed concrete or steel piles by means of full-scale testing for several simulated cofferdam
conditions, including (1) marine conditions, (2) fresh water conditions, and (3) drilling fluid
conditions. Normal pile surfaces and the situation of soil-caked piles were investigated.

FINDINGS
Both model and full-scale tests were carried out. In the model tests, a total of 36 one-third scale

specimens were testedtwenty-eight prestressed concrete and eight steel. Bonded embedment depth
in the seal slab was varied between d to 2d where d was the size of the pile. The results of these tests
indicated that shear stress variation was non-uniform, leading to larger computed bond stresses with
shallower embedment. Values were least for drilling fluid. Concrete piles had better bond with the
seal concrete than steel piles. Soil-caked condition was found to be relevant for the drilling fluid
situation only. In other cases, it was washed away from the pile surface.
In the full-scale tests, the 32 specimens tested were divided equally between steel and concrete. The
prestressed piles were 14 in. square and the steel piles were 14 in. deep wide flange sections.
Embedment depth, D, was varied between 0.5d to 2d ( i.e., 7 to 28 in., with the larger depth reserved
for the drilling fluid condition). Four of the sixteen prestressed piles were cast with embedded gages
located at the top, the middle, and the bottom of the interface region. The results of the full-scale
tests were similar to those from the one-third scale tests. The most important findings were (1) that
loads were transferred over a distance equal to the depth d of the pile, (2) that scale effects were
present--the average calculated bond stresses were lower for the full-scale tests than from the
corresponding scale model tests, (3) that prestressed piles cracked prior to bond failure, and (4) that
the seal slab cracked prior to bond failure.

CONCLUSIONS
Based on the test results it is proposed that the interface bond between piles and the seal slab be
restricted to an effective area in contact with the cast in place seal slab. The effective area is
calculated using the actual embedment depth (D) or the size of the pile (d), whichever is smaller.
The average bond stress over this region is limited to 300 psi for concrete piles and 150 psi for steel
piles. These values are reduced by a third (i.e., 100 psi and 50 psi, respectively) in cases where
drilling fluid is used. Application of the proposed values to the conditions related to the full-scale
tests led to average factors of safety in excess of two for both the prestressed and steel piles.
However, tension loads taken by the piles should not lead to cracking (concrete) or exceed the
allowable tension load (steel) of the piles. Nor should the seal slab crack.

This research project was conducted by and Gray Mullins, Ph.D, P.E. and Rajan Sen, Ph.D, P.E., at
the University of South Florida. For more information on the project, contact Moussa Issa, Ph.D,
P.E., S.E., at (850) 414-4291, moussa.issa@dot.state.fl.us